M4 Muscarinic Receptors Are Involved in Modulation of Neurotransmission at Synapses of Schaffer Collaterals on CA1 Hippocampal Neurons in Rats

Journal of Neuroscience Research 87:691–700 (2009)

Gonzalo Sańchez,1 Lucas de Oliveira Alvares,2,3 Marıá Victoria Oberholzer,1 Bruna Genro,2 Jorge Quillfeldt,2 Jaderson Costa da Costa,3 Carlos Cervenãnsky,4 Diana Jerusalinsky,1 and Edgar Kornisiuk1* 1Laboratorio de Neuroplasticidad y Neurotoxinas, Instituto de Biologıá Celular y Neurociencias, Facultad de Medicina, Universidad de Buenos Aires y CONICET, Buenos Aires, Argentina 2Laborato´rio de Psicobiologia e Neurocomputaçaõ, Dep. de Biofisica, Instituto de Biocieˆncias, Universidade Federal do Rio Grande do Sul, Porto Alegre, Brazil 3Laborato´rio de Neurocieˆncias, Instituto de Pesquisas Biome´dicas, Pontificia Universidade Cato´lica do Rio Grande do Sul, Porto Alegre, Brazil 4Instituto Pasteur de Montevideo e IIBCE, Montevideo, Uruguay

All five subtypes of muscarinic acetylcholine receptors Key words: muscarinic acetylcholine receptor; CA1 (mAChR; M1–M5) are expressed in the hippocampus, synapses; long-term potentiation; rat hippocampus; where they are involved both in cognitive functions and muscarinic toxin 3 in synaptic plasticity, such as long-term potentiation (LTP). Muscarinic toxins (MTs) are small proteins from All five subtypes of muscarinic acetylcholine recep- mamba snake venoms that display exquisite discrimi- tors (mAChR; M1–M5; Bonner et al., 1987) are nation between mAChRs. MT1 acts as an agonist at expressed in the hippocampus (see Volpicelli and Levey, M1 and an antagonist at M4 receptors, with similar 2004). mAChR of the hippocampus are considered to affinities for both. MT3, the most selective antagonist be involved in cognitive functions, because their activa- available for M4 receptors, infused into the CA1 region tion by nonselective agonists facilitates memory reten- immediately after training caused amnesia in the rat, tion, whereas antagonists produce amnesia (Bartus et al., indicating the participation of M4 receptors in memory 1982; Fibiger, 1991; see Jerusalinsky et al., 1997). They consolidation. Our goal was to investigate the parti- appear to be involved in synaptic plasticity such as long- cipation of M4 receptor in neurotransmission at the term potentiation (LTP), an increase in synaptic efficacy hippocampal Schaffer collaterals-CA1 synapses. Two proposed to underlie memory formation (Bliss and different preparations were used: 1) field potential Lomo, 1973). Accordingly, nonselective muscarinic ago- recordings in freshly prepared rat hippocampal slices nists enhance and antagonists disrupt LTP (Huerta and with high-frequency stimulation to induce potentiation Lisman, 1993; Ye et al., 2001; Leung et al., 2003; Li and 2) whole-cell voltage clamp in cultured hippocam- et al., 2007). The lack of ligands selective enough to dis- pal organotypic slices with paired stimuli. In preparation criminate between receptor subtypes has made it difficult 1, a dose of MT3 that was previously shown to cause to identify the physiological roles of particular subtypes. amnesia blocked LTP; the nonselective antagonist scopolamine blocked LTP without affecting basal transmission, although it was depressed with higher concen- The last two authors contributed equally to this work. tration. In preparation 2, basal transmission was Contract grant sponsor: University of Buenos Aires; Contract grant num- decreased and LTP induction was prevented by an ber: M040; Contract grant sponsor: CONICET; Contract grant number: MT3 concentration that would bind mainly to M4 recep- PIP6086; Contract grant sponsor: FONCyT; Contract grant number: PICT05-14346. tors. Although M1 receptors appeared to modulate transmission positively at these excitatory synapses, *Correspondence to: Edgar Kornisiuk. Lab. Neuroplasticidad y Neuro- M1 activation concomitant with M4 blockade (by MT1) toxinas, Instituto de Biologıá Celular e Neurociencias, Fac. Med., Univ. only allowed a brief, short-term potentiation. Accord- de Buenos Aires, 2155 Paraguay st., 2nd floor, 1121 Buenos Aires, Argentina. E-mail: [email protected] ingly, M4 blockade by MT3 strongly supports a permissive role of M4 receptors and suggests their necessary Received 28 May 2008; Revised 26 June 2008; Accepted 11 July 2008 participation in synaptic plasticity at these synapses. Published online 24 September 2008 in Wiley InterScience (www. VC 2008 Wiley-Liss, Inc. interscience.wiley.com). DOI: 10.1002/jnr.21876

' 2008 Wiley-Liss, Inc. 692 Sa´nchez et al.

The use of knockout mice hinted at a nonessential most contradictory aspects concerning HFS is its doubt- modulatory contribution of M1 (Miyakawa et al., 2001) ful physiological relevance, insofar as there appears not and M3 receptors (Shinoe et al., 2005) to learning and to be any equivalent activity in vivo. Therefore, we memory and to LTP. In M1–/– mice, there was a mild decided to use the pairing protocol for LTP induction in reduction in theta burst stimulation LTP (TBS-LTP) at our whole-cell experiments because this kind of stimula- the Schaffer collateral-CA1 synapse, but there were no tion appears to be similar to the activity that is going on changes on high-frequency stimulation LTP (HFS-LTP), in animals in behavioral assays. with slight impairments in learning (Anagnostaras et al., The results suggest that both M1 and M4 receptors 2003). M2 and M4 are both auto- and heteroreceptors are positively involved in transmission, with different that couple to Gi proteins and share some ligand binding effects in potentiation at these hippocampal synapses. We properties, making it difficult to discriminate between propose that M4 receptors have a permissive role in them. The current M2/M4 antagonists appeared to transmission and suggest their essential participation in improve performance in some behavioral tasks (Quirion synaptic plasticity at these excitatory synapses. et al., 1995; Rowe et al., 2003) and to enhance consolidation (Baratti et al., 1993). However, it was recently MATERIALS AND METHODS reported that the relatively selective M2 antagonist AF- DX-116 injected into the hippocampus produced a Muscarinic toxins MT1 and MT3 were purified from trend to improve acquisition, although it did not affect Dendroaspis angusticeps snake venom (J. Leakey Ltd., Kenya, long-term memory (Herrera-Morales et al., 2007). Auto- East Africa; Jerusalinsky et al., 1992). [3H]N-methylscopol- inhibition of acetylcholine (ACh) release in mouse hip- amine (84 Ci/mmol) was purchased from Dupont-New Eng- pocampus and cerebral cortex would be mediated mainly land Nuclear (Boston, MA). by M2 receptors (Zhang et al., 2002). However, M2–/– We used adult male Wistar rats from the School of Vet- mice showed a decrease in LTP amplitude and deficits erinary Sciences. Experiments with rats were performed in in working memory (Seeger et al., 2004). strict accordance with the Review Committee of the School Basal locomotor activity was slightly increased in of Veterinary Sciences, University of Buenos Aires, the Brazil- M4–/– mice (Gomeza et al., 1999), and it was suggested ian law for the recommendations of the Brazilian Society for that this receptor could be involved in modulation of Neurosciences, and the International Brain Research Organi- attention (Felder et al., 2001). However, there are no zation (IBRO) and are in compliance with the National Insti- reports on either learning and memory or LTP in tutes of Health Guide for care and use of laboratory animals (pub- M4–/– mice and only one recent report on pharmaco- lication No. 85-23, revised 1985). logical studies of synaptic plasticity with M4 selective agents (Shirey et al., 2008). Muscarinic toxins (MTs) are small proteins from Radioligand Binding Assays mamba snake venoms that display exquisite discrimina- Hippocampi from adult Wistar rats (200–250 g) were tion among mAChRs. MT1 acts as an agonist at M1 and homogenized in 10 volumes (w/v) of ice-cold hypotonic an antagonist at M4 receptors (Jerusalinsky et al., 1995; buffer (10 mM HEPES/NaOH, 0.3 mM EGTA, 2.3 mM Kornisiuk et al., 2001), with similar affinities for both MgCl2, pH 7.4) plus 0.32 M sucrose, then centrifuged for (Kornisiuk et al., 1995). MT3 is the most selective an- 10 min at 1,000g at 48C. The supernatant was centrifuged at tagonist available for M4 receptors (Max et al., 1993; Jer- 11,000g for 20 min at 48C. The pellet was resuspended, incu- usalinsky et al., 1998). Both toxins show negligible bind- bated for 20 min in 20 ml hypotonic buffer, and centrifuged ing to the other receptors. MT1 infused into the CA1 at 100,000g for 45 min at 48C. The final pellet was resus- 3– 1 1 region of rat hippocampus immediately after training pended in phosphate buffer (PO4 ,Na,K 50 mM, pH facilitated memory retention (Jerusalinsky et al., 1995), 7.4) to a protein concentration of 1.5 6 0.5 mg/ml, deter- whereas MT3 caused amnesia, indicating the participa- mined according to Bradford (1976). tion of M4 receptors in memory consolidation (Jerusa- Inhibition experiments were performed with 0.5-ml ali- linsky et al., 1998; Ferreira et al., 2003). quots of 0.1 mg protein/ml membranes in phosphate buffer, The main goal of this study was to investigate the carried out in triplicate. MTs concentrations were estimated role of M4 receptor in hippocampal neurotransmission from dilution of a stock solution, by their specific absorbance 0.1% 3 and synaptic plasticity. Two different preparations and (A276 nm is 2.21 for MT1 and 1.47 for MT3). H-N- stimulation protocols were used: field potential record- methyl scopolamine (3H-NMS) was added (at 0.5 nM final ings in rat hippocampal slices with high-frequency stim- concentration), and the aliquots were incubated for 1 hr at ulation (HFS) and whole-cell voltage clamp for record- 378C. Similar aliquots were incubated with the addition of ings from pyramidal cells in cultured hippocampal orga- 10 lM atropine for nonspecific binding. The incubation was notypic slices, with paired stimuli. stopped with ice-cold diluted phosphate buffer, followed by HFS is the most typical protocol employed for rapid filtration through glass fiber filters (Whatman GF/B) in LTP induction, and most of the pharmacological data on a Millipore apparatus connected to a vacuum pump. Filters LTP have been produced with HFS in field recordings. were dried at 708C for 2 days, counted for radioactivity in a This is the main reason why we decided to use this pro- liquid scintillation counter (scintillating cocktail: 2,5-diphenyl- tocol in our experiments. On the other hand, one of the oxazole/xylene 5 g/liter).

Journal of Neuroscience Research M4 Receptor Modulation of CA1 Synapses 693

Field Recordings in Acute Slices were evoked with monophasic voltage pulses (1–10 V, 200 l Adult male Wistar rats (250–300 g) were anesthetized sec) at 0.3 Hz, and membrane potential was clamped at –60 with thiopenthal (50 mg/kg, i.p.) and decapitated. Their mV. In similar assays where membrane potential was set at – brains were quickly removed and immersed in ice-cold low- 60 mV, Barria and Malinow (2002) showed that 90% of the 1 Ca2 artificial cerebrospinal fluid (ACSF, in mM: 130 NaCl, evoked excitatory postsynaptic current (EPSC) corresponds to ions permeating through AMPA-R and that only 5% or less 3.5 KCl, 1.3 NaH2PO4, 5 MgCl2, 0.2 CaCl2, 10 D-glucose, corresponds to N-methyl-D-aspartate receptor (NMDA-R)- 24 NaHCO3, pH 7.3–7.4, by gassing with 95/5% O2/CO2). Transverse hippocampal slices (400 lm thick) were obtained mediated currents. A few minutes after gaining whole-cell using a vibrating tissue slicer (Vibroslice 725 M; Campden access, LTP was induced through a paired-stimuli protocol, Instruments) and transferred to a holding immersion-type i.e., stimulation frequency at 3 Hz and membrane potential at 1 chamber at room temperature in normal Ca2 ACSF (ACSF, 0 mV, during 2 min. After induction, stimulation was restarted at basal conditions. Stock solutions of 0.1 mM MTs in mM: 130 NaCl, 3.5 KCl, 1.3 NaH2PO4, 2 MgCl2,2 were prepared in ACSF and stored at –208C. Picrotoxin 0.1 CaCl2, 10 D-glucose, 24 NaHCO3, pH 7.3–7.4, by gassing mM was included in circulating ACSF during recordings. with 95/5% O2/CO2). Slices were allowed to recover for at least 90 min and transferred to a recording interface-type chamber, perfused at 2–3 ml/min with ACSF with normal 1 Data Acquisition and Analysis Ca2 . Binding data were analyzed by nonlinear regression Standard extracellular electrophysiology techniques were using GraphPad Prism version 4.00 for Windows (GraphPad used to record field excitatory postsynaptic potentials (fEPSPs) Software, San Diego, CA; www.graphpad.com). Two differ- from the dendritic region of CA1 neurons (stratum radiatum) ent equations for either one-site or two-site model (sigmoidal in response to stimulation of the Schaffer’s collaterals afferent dose-response curves) were fitted to data from binding assays; pathway, using square current pulses (60–120 lA, 0.2 msec, the software was used to compare the results to determine the 0.05 Hz; Master 8; AMPI, Israel). The stimulation electrode best regression according to F-test by balancing the change in consisted of a twisted bipolar pair of 75-lm platinum-iridium sum of squares and the degrees of freedom for each experi- wires (A-M Systems). Recording electrodes were pulled on a mental data set. For the two-site regression model, two per- horizontal micropipette puller (Sutter P-87; Sutter Instrument) centages were calculated, one for the proportion of high-affin- from borosilicate glass capillaries filled with 0.9% NaCl (elec- ity sites and another for the low-affinity sites; and there were trode resistance 0.5–10 MX). After a stable baseline-evoked two EC50 values, one for high- and another for low-affinity response was observed, the HFS protocol was applied (four populations. trains of 1 sec duration at 100 Hz, pulse duration of 0.2 msec, In acute hippocampal field recordings, generated data with an intertrain interval of 20 sec). Field potentials were were amplified 1,0003 and low-pass filtered at 0.6 kHz monitored for at least 60 min after the HFS. (CyberAmp 320; Axon Instruments, Foster City, CA), digi- A micropippete containing either MT3 or scopolamine tized (Digidata; Axon Instruments), and recorded (Axo-Clamp diluted in ACSF was placed next to the stratum radiatum, and 2B; Axon Instruments). In whole-cell experiments, recordings the drug or its diluent (vehicle) was ejected by pressure pulse were made with an Axopatch-1D amplifier (Axon Instru- (4 ll; named puff) generated with a pneumatic pump (PV830 ments). Junction potentials were not corrected. Pneumatic Pico Pump; WPI, as in Salamoni et al., 2005) onto Parameters from electrophysiology recordings were the recording region, 2 min before LTP induction. obtained through Clampfit 9.2 (Axon Instruments) analysis: amplitude of evoked currents was taken from EPSCs and Whole-Cell Recordings in Organotypic Cultures potential’s mean slope from fEPSPs. Statistical analysis was 6 Hippocampal organotypic culture slices (400 lm) were performed with GraphPad Prism; all data refer to mean prepared from P6–P7 Sprague Dawley rats using a tissue SEM. An exponential association or decay model was adjusted chopper (Stoppini et al., 1991). After 6–7 days in vitro, slices to basal evoked transmission results. were transferred to a recording chamber continuously perfused with artificial cerebrospinal fluid (ACSF in mM: 119 NaCl, RESULTS 2.5 KCl, 4 CaCl2, 4 MgCl2, 26 NaHCO3, 1 NaH2PO4,11 3 glucose, and 0.001 2-chloroadenosine, pH 7.4). A volume of Inhibition Curves of H-NMS Specific 50 ml of ACSF driven by a peristaltic pump was bubbled Binding by MTs 3 with 95/5% O2/CO2 at 22–248C in a closed circuit (2–3 ml/ Inhibition curves of the binding of H-NMS mus- min). Patch pipettes were filled with standard intracellular so- carinic antagonist by either MT1 or MT3, in synaptoso- lution containing (in mM): 115 cesium methanosulphonate, mal membranes from the hippocampal formation, are 20 CsCl, 10 HEPES, 2.5 MgCl2,4Na2ATP, 0.4 Na3GTP, shown in Figure 1. The inhibition by MT1 was better 10 sodium phosphocreatine, 0.6 EGTA, pH 7.25; the elec- fitted to a one-site model, with a maximal inhibition of trode resistance was 4–6 MX. Whole-cell voltage-clamp 83.2% 6 2.2% and with an IC50 5 171.6 nM (Ki 45.9 recordings were obtained from CA1 pyramidal cells under mi- nM). For the inhibition by MT3, the curve was better croscopic guidance. A bipolar electrode (CE2C55; Frederick fitted to a two-site model; high-affinity sites correspond Haer, Bowdoinham, ME) was placed on Schaffer collaterals, to 24.3% 6 8.5% of the total sites, with an IC501 of 250 lm away from the soma of the recorded cell; responses 0.85 nM (Ki1 0.23 nM), whereas the total inhibition

Journal of Neuroscience Research 694 Sa´nchez et al.

Whole-Cell Voltage-Clamp Recordings in Organotypic Cultures Whole-cell voltage clamp of hippocampal pyramidal neurons in organotypic cultures was used to explore the role of mAChR in modulating the activity at Schaffer collateral-CA1 glutamatergic synapse. EPSCs were evoked by stimulation at 0.3 Hz throughout the whole assay and the holding potential was set at – 60 mV. Once the recording appeared stable for at least 5 min, MTs were included in the circulating ACSF to a final concentration of 100 nM, and recordings were followed for another 20 min. Figure 3 depicts the normal- ized amplitudes of EPSCs, before and after perfusion of 100 nM MT1 or MT3. The curve for each MT fitted to a single exponential model. The steady states reached and the time constants were estimated: there was a 38.3% 6 1.8% increase in EPSC amplitude after MT1 (tau 5 1.56 min; n 5 6; Fig. 3A) and a 54.2% 6 2.1% 3 5 5 Fig. 1. Inhibition curves with MT1 and MT3 of H-NMS specific reduction after MT3 (tau 3.33 min; n 6; Fig. 3B). binding in hippocampal membranes. Inhibition by MT1 (solid circles, Taking into account the pharmacological profile of MT3 n 5 4) fit better to a one-site competition curve, with Ki of 45.9 nM as a selective M4 antagonist and that of MT1 as agonist 5 6 (IC50 171.6 nM), and a bottom of 16.8% 2.2%. Inhibition by MT3 at M1 and antagonist at M4, the results indicate that 5 (open circles, n 4) fit better to a two-site competition model, with Ki1 both receptors appear to be involved in modulating 5 5 of 0.23 nM (IC501 0.85 nM) and Ki2 of 61.8 nM (IC502 226.7 transmission at these synapses. nM), bottom of 15.2% 6 6.7% and of 24.3% 6 8.5% for MT3 high-af- To evaluate the participation of mAChR in LTP finity sites. induction, potentiation was induced by stimulation at 3 Hz for 2 min to the Schaffer collaterals, in conjunction with postsynaptic depolarization to 0 mV. After this reached 84.8% 6 6.7%, and the IC502 was 226.7 nM stimulation, holding potential was turned back to –60 (Ki2 61.8 nM). mV, and the frequency of stimulation was reduced to 0.3 Hz. EPSC amplitudes without the toxins and with either MT1 or MT3 are shown in Figure 4. In control Field Recordings in Hippocampal Slices recordings, the EPSCs were significantly increased immediately after the induction protocol, and stabilized af- Field potentials were recorded in freshly prepared ter 20 min at 100% above the basal level for at least hippocampal slices from rat brain. Each slice received ei- another 20 min. In the presence of either MT1 or MT3 ther one of two different doses of MT3, or one of sco- (100 nM), stimulation with the LTP-induction protocol polamine, or the vehicle administered as a puff to the was followed by an immediate increase in evoked CA1 region; 2 min later, the Schaffer collaterals were EPSCs, although it did not reach the levels of potentia- stimulated with an HFS protocol to induce potentiation tion found in the control experiments (Fig. 4A). How- (Fig. 2). Figure 2A shows the fEPSP slope over the 60- ever, with MT3, the evoked EPSC returned to basal min time course of the experiments. In control assays level in about 6 min, whereas, in the case of MT1, the (vehicle only), HFS produced an immediate and robust enhancement lasted for 15–18 min and then decayed increase in evoked responses to about 50% above the ba- until it was not different from basal levels at about sal level. Then, the EPSP slope increased to about 100% 20 min (Fig. 4B). Therefore, the effects with both of the above control levels in 30–40 min, and this potentiation toxins followed different temporal courses and patterns. persisted during the whole recording (60 min). How- In summary, some degree of potentiation appeared to ever, after a puff of 0.4 lg/ll MT3, HFS did not pro- take place but did not last as persistent potentiation in duce an immediate increase in fEPSPs (not shown), the presence of either MT, suggesting that blockade of although a potentiation appeared after 10 min. During M4 receptors is sufficient to suppress LTP expression, the last 10 min of the recording, this potentiation even though M is activated. was not statistically different from potentiation levels ob- 1 served under control conditions (Fig. 2B). After administration of 4 lg/ll of MT3, there was no evidence of DISCUSSION potentiation with HFS, insofar as the fEPSPs slope remained at basal levels (Fig. 2A,B). After the delivery of Hippocampal mAChR Subtypes 4 lg/ll scopolamine through a puff, there was no Muscarinic toxins are small proteins from green potentiation; furthermore, the fEPSPs slope was reduced mamba snake venom that display exquisite discrimina- to about 50% (on average) of its basal level. tion among mAChR subtypes. The different selectivity

Journal of Neuroscience Research M4 Receptor Modulation of CA1 Synapses 695

Fig. 2. LTP blockade by acute administration of MT3 or scopola- gray) and at the time indicated by the horizontal bar (black; A, right). mine. A: Left: LTP was induced by the high-frequency stimulation Scale bars 5 5 mV, 10 msec. B: Scatterplot showing the mean slope protocol (HFS, arrow). Two minutes before the induction, vehicle values for 10 min baseline and those corresponding to the last 10 (circles, n 5 6), 4 lg/ll MT3 (open triangles, n 5 5), 0.4 lg/ll min of recording indicated by the black horizontal bar in A. Means MT3 (solid triangles, n 5 5, B), or 4 lg/ll scopolamine (lozenges, n from these data were used for statistical analysis. *Signiﬁcant differen- 5 5) was delivered through a puff aimed at the recording zone. Data ces between baseline and last 10 min of recording for each group are mean of fEPSP slope 6 SEM. Right: Sample of fEPSP single (P < 0.01, paired Student’s t-test). traces corresponding to basal conditions prior to MTs exposure (in

3 profiles of MT1 (similar affinities for M1 and M4) and from autoradiograms of the displacement of H-NMS MT3 (highest affinity for M4, followed by that for M1) binding to brain slices by classical antagonists or by MTs allowed an estimate to be made of the proportions of (Jerusalinsky et al., 2000) and autoradiograms of the 125 M1 and M4 mAChRs in the hippocampus. From the binding of I-MT3 (Adem and Karlsson, 1997) and curves of the inhibition of 3H-NMS specific binding by those with biotinylated MT3 (Santiago and Potter, MT3, we have estimated that non-M4 receptors, mainly 2001). The estimated IC50s are in agreement with the involving the M1 receptor subtype, would amount to previously reported affinities of the MTs; i.e., the esti- about 60% of total mAChR, whereas the M4 subtype mated affinity of MT3 at the M4 receptor was 266-fold would be 24% in this structure (Fig. 1). These values are higher than at M1, whereas MT1 did not differentiate in agreement with those here reported for MT1 and between the two receptor subtypes (Jerusalinsky and with previous reports of partial estimation of mAChR Harvey, 1994).

Journal of Neuroscience Research 696 Sa´nchez et al.

Fig. 4. LTP induction in the presence of muscarinic toxins. A: LTP was induced by the paired protocol (arrow) either in the absence (solid circles) or in the presence of MT1 (open circles) or MT3 (triangles). Each point represents mean of EPSC amplitudes 6 SEM calculated from five (MT1, MT3) or eight (control) independent experiments. B: Single traces of EPSCs prior to paired protocol (basal, in gray) and at the time window indicated by the horizontal bar Fig. 3. Effect of muscarinic toxins on evoked basal transmission. Ba- 5 sal evoked responses were monitored for 20 min. Either MT1 (A)or in A (induced, in black). Scale bars 10 pA, 10 msec. C: Scatterplot MT3 (B) was present in the circulating ACSF from min 5 onward, showing the amplitude mean within the time window indicated by the black horizontal bar in A. Means from these data were used for as indicated by horizontal bars. Dotted line corresponds to the < graphic representation of an exponential model fitted to both sets of statistical analysis. *Mean significantly different from 1 (P 0.05, experimental data (mean 6 ESM, n 5 6); best-fit parameter values one-sample Student’s t-test; 95% confidence interval for each mean were: for MT1, plateau 1.38 6 0.01 and tau 1.56 6 0.09 min; for was 1.978–1.800 for control, 1.059–0.812 with MT1, and 1.009– MT3, plateau 0.46 6 0.01 and tau 3.33 6 0.02 min. Scale bars for 0.762 with MT3). single EPSCs 5 10 pA, 10 msec.

well as in GABAergic and glutamatergic terminals, which would be involved in control of neurotransmitter Use of selective antibodies has shown that M2 is widely expressed in the CNS and in the periphery, release, there is also immunological evidence of a postsynaptic localization at least in the dentate gyrus (Rouse whereas M4 is preferentially expressed in the forebrain (Vilaro´ et al., 1993; see also Volpicelli and Levey, 2004). et al., 1999). Both subtypes appeared to act mainly as both presynaptic auto- and heteroreceptors. Zhang et al. (2002) have mAChR and Neurotransmission at CA1 reported that autoinhibition of ACh release is mediated The precise signaling pathways through which mainly by M2 receptors in the mouse hippocampus and native mAChR subtypes exert modulation of neuronal cerebral cortex and by M4 receptors in the striatum. activity in the hippocampus has remained elusive because In agreement with other reports using different of the diffuse cholinergic innervation there (Descarries approaches, our results show that there is a conspicuous et al., 1997) and because of the lack of ligands selective expression of mAChR, mainly M1, in the hippocampus enough to discriminate between receptor subtypes (see and that more than 80% corresponds to M1 and M4 sub- Alexander and Peters, 2000). Miyakawa et al. (2001) types. Furthermore, Levey et al. (1995) have previously reported that M1–/– mice performed as well as their shown that neurons in the hippocampus, i.e., interneur- wild-type (WT) littermates in various hippocampus-de- ons and pyramidal and granule cells, are immunopositive pendent tasks but showed some deﬁcits in other tasks for M1 and M4 receptors, with a weak M2 immunostain- that correlated with the degree of hyperactivity dis- ing in these cells (Levey et al., 1995). In spite of a pre- played. Anagnostaras et al. (2003) reported a mild reduc- dominant presynaptic localization of M4 in cholinergic as tion of TBS-LTP at Schaffer collateral-CA1 synapses in

Journal of Neuroscience Research M4 Receptor Modulation of CA1 Synapses 697

M1–/– mice, although there were no changes in HFS- CA1, we intended to mimic the in vivo injection. At LTP. On the other hand, there were mild and task- the dose of MT3 that had no effect on memory reten- selective impairments in learning. In addition, the treat- tion, LTP was preserved (Fig. 2B), whereas, at the dose ment of both WT and M1–/– mice with the classical that caused amnesia (Jerusalinsky et al., 1998; Ferreira nonselective antagonist scopolamine resulted in similar et al., 2003), LTP was blocked (Fig. 2A). Indeed, with cognitive deficits, suggesting that M1 does not play criti- the lower dose, the potentiation appeared to be delayed, cal roles in learning and memory. Furthermore, Shinoe but it finally reached control levels. We do not have an et al. (2005) detected an intact hippocampal HFS-LTP explanation for this, but it appears to be an interesting both in M1–/– and in M3–/– mice; however, M1 case of delayed plasticity, either by temporarily decreas- appeared to be involved in LTP enhancement by the ing transmission or by affecting potentiation mechanisms, agonist carbachol (50 nM). In M2–/– mice, TBS-LTP which deserves further investigation. Because MT3 binds was not abolished, although the amplitude of the poten- reversibly to M4 receptors (Max et al., 1993; Olianas tiation was significantly reduced in the hippocampus, et al., 1996; see Bradley, 2000) and the slices were con- and the mice also showed deficits in working memory tinuously perfused with buffer during and after the puff, (Seeger et al., 2004). On the other hand, it was previ- the toxin was washed out, although with the higher ously reported that currently available M2/M4 antago- dose the effect persisted. nists improved cognitive performance in WT mice and According to Sokolov and Kleschevnikov (1995), rats (Baratti et al., 1993; Quirion et al., 1995; Rowe the drug concentration at the slice after delivery through et al., 2003). Herrera-Morales et al. (2007) reported that the puff would be about 100-fold lower than the con- direct application of the relatively selective M2 antagonist centration in the micropipette; furthermore, the washout AF-DX-116 into the dorsal hippocampus did not affect begins immediately. Thus, the estimated concentration acquisition, memory formation, or even long-term of scopolamine that reduced basal transmission level (Fig. memory. On the other hand, pirenzepine did not disrupt 2A) would initially be on the order of 10–4 M, whereas acquisition but seriously impaired long-term memory. MT3 would be on the order of 10–6 M. Although there This effect of pirenzepine was attributed to M1 block- was a prolonged depression after scopolamine, which ade. However, pirenzepine shows similar overlapping persisted while it was being washed out and hence could ranges of affinities for both M1 and M4 receptors (see be interpreted as LTD, the antagonist concentration used Alexander and Peters, 2000). was rather high (10 mM as a puff, likely equivalent to M4–/– mice display a small increase in basal loco- 100 lM in situ) to be considered truly specific. Only motor activity (Gomeza et al., 1999), and it was sug- with this high concentration was there a clear and persis- gested that this receptor could play a role in attention tent reduction in transmission, but we have not yet car- (Felder et al., 2001). Zhang et al. (2002) suggested a role ried out further studies to investigate the underlying for M4 receptor in the control of transmitter release. mechanisms. Furthermore, there are no reports on LTD However, to our knowledge, there are no reports on produced either by nonselective or by subtype-selective learning and memory tests with M4–/– mice or on hip- antagonists. The ‘‘chemical LTD’’ was always reported pocampal LTP. Furthermore, there was no previous to be produced by addition of agonists. We have already report regarding the role of M4 receptors in hippocampal begun to develop assays to clarify further the effect of LTP. scopolamine. In a set of preliminary experiments (n 5 Pharmacological assays with MTs showed that the 3), where scopolamine was continuously perfused and blockade of M4 hippocampal receptors resulted in amne- hence the concentration was accurately controlled, there sia in rats. The infusion of MT3 (4 lg/ll) into the CA1 was not reduction neither in basal transmission nor in region of the hippocampus immediately after training induced potentiation up to 5 lM, whereas, at either 25 produced amnesia (Jerusalinsky et al., 1998; Ferreira or 50 lM scopolamine, there was a blockade of LTP et al., 2003), indicating the participation of M4 receptors induction without modification in basal transmission in memory consolidation. On the other hand, MT1 (data not shown). These results are in agreement with facilitates retention, suggesting that M1 postsynaptic re- data reported by Ye et al. (2001), showing that 10 lM ceptor activation would predominate over other musca- scopolamine did not affect basal level of transmission, rinic influences, at least immediately after training (Jeru- although it significantly decreased the tetanus-induced salinsky et al., 1995). LTP at the same synapse. In this study, we used MT3, the most selective an- In the whole-cell voltage-clamp recordings, basal tagonist for M4 mAChR known (Max et al., 1993; Jeru- transmission was decreased and induction of LTP by salinsky et al., 1998), to study the putative role of this paired stimuli was prevented by a concentration of MT3 receptor in hippocampal neurotransmission and synaptic that would bind mainly to M4 receptors (Jerusalinsky plasticity. We also used MT1, with similar affinities for et al., 1997; Figs. 3B, 4). Although we could not rule M1 and M4 and negligible binding to the other receptors out the possibility that this decrease by MT3 was a last- (Kornisiuk et al., 1995). ing depression, this could not be concluded from those The use of field recordings allowed us to adminis- experiments because the toxin was always present. On ter doses similar to those employed in behavioral experi- the other hand, there was an enhancement of basal trans- ments, and, by delivering them locally through a puff to mission with MT1, which might be due to its agonistic

Journal of Neuroscience Research 698 Sańchez et al. activity at M1 receptor, but this toxin also prevented far as our whole-cell assays were performed in the pres- LTP. Hence, the blockade of LTP by both MTs likely ence of picrotoxin, at least the participation of GABAA was due to their antagonism at M4 receptors, whereas control of transmission could be excluded. It is plausible the different effect during the first 20 min after induc- that muscarinic receptors at CA1 modulate several potas- tion of potentiation could be explained by the differ- sium conductances (IAHP,Im, and Ileak; see Halliwell, ential action of MTs at M1 receptors. Although M1 1990), the permeability of voltage-sensitive calcium receptors appeared to modulate transmission positively at channels (Tai et al., 2006), and several ligand-gated these synapses, M1 activation concomitant with M4 receptors, including the NMDA receptor (Markram and blockade (by MT1) allowed only a brief short-term Segal, 1990). Although M4 receptors might be involved potentiation but was not enough to overcome M4 in some of those functions, there is no evidence to cor- blockade. Because the M1 receptor is the most abundant roborate their participation in any of them. and MT1 shows higher affinity for M1 than for M4, the Taking into account the results from whole-cell effect of the toxin could have been explained on the ba- experiments, we can suggest that, under basal conditions, sis of a differential affinity for the neurotransmitter ACh. when NMDA currents were suppressed by setting the However, ACh has rather similar affinities for M1–M4 membrane potential at –60 mV, M4 would positively receptors (Jakubı´k et al., 1998). The decrease in basal modulate AMPA currents rather than being involved in transmission associated with M4 blockade suggests a per- the muscarinic depression observed in CA3–CA1 and missive and necessary role of M4 receptors at these exci- other glutamatergic synapses (Sim and Griffith, 1996; tatory synapses. Because there was a concentration of Yajeya et al., 2000; Atzori et al., 2005). After the induc- scopolamine that was able to block LTP without signifi- tion protocol, NMDA receptors would be recruited and cantly affecting basal transmission or producing a lasting could be involved in muscarinic M4 modulation, depression, and even an increase in basal transmission by although this is rather speculative, and further research is M1 activation could not prevent suppression of LTP by necessary to clarify the point. M4 receptor blockade, we can strongly suggest the par- It has been assumed that both types of induction ticipation of M4 receptor in synaptic plasticity beyond protocols used here share the main mechanisms underly- the modulation of basal transmission. ing the plastic change, including NMDA receptor de- The LTP reported here is likely to be the ‘‘classical’’ pendence. Nevertheless, recent investigation points to N-methyl-D-aspartate (NMDA)-dependent monosynap- the possibility that a huge diversity of cascades is differ- tic LTP, insofar as it was induced by presynaptic stimulation entially triggered by each protocol, yet leading to similar coincident with postsynaptic depolarization (by setting results (Lynch et al., 2007). If this were the case, our membrane potential at 0 mV). Thus, the effect of M1 re- results point toward an essential role of M4 in synaptic ceptor activation shown here would not appear directly plasticity, in that its integral function appears to be nec- related to that caused by a puff of ACh at the apical den- essary in both paradigms. drites of CA1 pyramidal cells eliciting an LTPIP3 as Recently, Shirey et al. (2008) showed that a selec- recently reported by Fernańdez de Sevilla et al. (2008). tive allosteric agonist (20j) for M4 receptors increased This LTPIP3 does not depend on NMDA receptors and carbachol-induced depression of excitatory glutamatergic does not need pre- and postsynaptic correlated action transmission at the same synapse (carbachol produces op- potentials; furthermore, the LTPIP3 would not be affected posite effects depending on the concentration used; by muscarinic presynaptic receptor action and would be Auerbach and Segal, 1996), although it did not show mediated exclusively through postsynaptic mechanisms. any effect at inhibitory transmission. Their result appears Those authors found that both types of LTP tended to add contradictory to ours, although it is difficult to compare linearly and suggested that both mechanims could supply the two because the conditions of the assays were rather different functions in the same neurons. Here we reported different. In our case, there was electrical stimulation that both pre- and postsynaptic mAChR could directly putatively activating endogenous release of neurotrans- modulate the ‘‘classical LTP’’ that depends on pre- and mitters, with or without the application of the M4 postsynaptic mechanisms. antagonist, which appeared to depress glutamatergic Muscarinic transmission in the hippocampus has transmission. Instead, the authors added the agonist car- been related to many different cellular functions, ranging bachol in concentrations known to cause transient from modulation of a variety of ionic currents to a wide depression, and this effect was enhanced by M4 activa- span of biochemical signaling in pyramidal neurons tion. It could be speculated that different mAChR through both direct and indirect biochemical interactions populations would be recruited in each case. When (see Cobb and Davies, 2005). However, information on Fernańdez de Sevilla et al. (2008) applied a tetanic stim- the role of each receptor subtype is scarce; in particular, ulation to the cholinergic afference to the hippocampus there is no information on M4 participation. If we in vivo, the evoked fEPSP slope by Schaffer collaterals accept a predominatly presynaptic localization in cholin- stimulation significantly increased (210%) in a stable and ergic as well as in GABAergic and glutamatergic termi- lasting manner (LTP). Therefore, this result should be nals, M4 activation would inhibit neurotransmitter interpreted as the physiological effect of endogenous release but would be able to enhance excitability indi- ACh released from medial septum neurons into the rectly, i.e., by inhibiting GABA release. However, inso- hippocampus.

Journal of Neuroscience Research M4 Receptor Modulation of CA1 Synapses 699

To the best of our knowledge, the paper by Shirey Gomeza J, Zhang L, Kostenis E, Felder C, Bymaster F, Brodkin J, Shan- et al. (2008) and the present paper constitute the first non H, Xia B, Deng C, Wess J. 1999. Enhancement of D1 dopamine receptor-mediated locomotor stimulation in M muscarinic acetylcho- indications regarding M4 receptor participation in neuro- 4 transmission and synaptic plasticity in the hippocampus, line receptor knockout mice. Proc Natl Acad Sci U S A 96:10483– respectively. These results together with the amnesia fol- 10488. Halliwell JV. 1990. Physiological mechanisms of cholinergic action in the lowing M4 blockade in the hippocampus point to a rele- hippocampus. Prog Brain Res 84:255–272. vant role of this receptor in cognitive functions and syn- Herrera-Morales W, Mar I, Serrano B, Bermu´dez-Rattoni F. 2007. Acti- aptic plasticity. vation of hippocampal postsynaptic muscarinic receptors is involved in ACKNOWLEDGMENTS long-term spatial memory formation. Eur J Neurosci 25:1581–1588. We are grateful to Francisco Urbano for his help Huerta PT, Lisman JE. 1993. Heightened synaptic plasticity of hippocampal CA1 neurons during a cholinergically induced rhythmic state. Na- and advice on field recordings. ture 364:723–725. Jakubı´kJ,Ba˜ca´kova´ L, El-Fakahany EE, Tueèk S. 1997. Positive cooper- REFERENCES ativity of acetylcholine and other agonists with allosteric ligands on Adem A, Karlsson E. 1997. Muscarinic receptor subtype selective toxins. muscarinic acetylcholine receptors. Mol Pharmacol 52:172–179. Life Sci 60:1069–1076. Jerusalinsky D, Harvey AL. 1994. Toxins from mamba venoms: small Alexander SPH, Peters JA, editors. 2000. TiPS receptor and ion channel proteins with selectivities for different subtypes of muscarinic acetylcho- nomenclature supplement, vol 11. New York: Elsevier. p 6–7. line receptors. Trends Pharmacol Sci 15:424–430. Anagnostaras SG, Murphy GG, Hamilton SE, Mitchell SL, Rahnama Jerusalinsky D, Cervenãnsky C, Penã C, Raskovsky S, Dajas F. 1992. NP, Nathanson NM, Silva AJ. 2003. Selective cognitive dysfunction in Two polypeptides from Dendroaspis angusticeps venom selectively in- acetylcholine M1 muscarinic receptor mutant mice. Nat Neurosci 6:51– hibit the binding of central muscarinic cholinergic receptor ligands. 58. Neurochem Int 20:237–246. Atzori M, Kanold PO, Pineda JC, Flores-Hernandez J, Paz RD. 2005. Jerusalinsky D, Kornisiuk E, Bernabeu R, Izquierdo I, Cervenãnsky C. Dopamine prevents muscarinic-induced decrease of glutamate release in 1995. Muscarinic toxins from the venom of Dendroaspis snakes with the auditory cortex. Neuroscience 134:1153–1165. agonist-like actions. Toxicon 33:389–397. Auerbach JM, Segal M. 1996. Muscarinic receptors mediating depression Jerusalinsky D, Kornisiuk E, Izquierdo I. 1997. Cholinergic neurotrans- and long-term potentiation in rat hippocampus. J Physiol 492:479–493. mission and synaptic plasticity concerning memory processing. Neuro- Baratti CM, Opezzo JW, Kopf SR. 1993. Facilitation of memory storage chem Res 22:507–515. by the acetylcholine-M2 muscarinic receptors antagonist AF-DX 116. Jerusalinsky D, Kornisiuk E, Alfaro P, Quillfeldt J, Alonso M, Verde ER, Behav Neural Biol 60:69–74. Cervenãnsky C, Harvey A. 1998. Muscarinic toxin selective for m4 Barria A, Malinow R. 2002. Subunit-specific NMDA receptor trafficking receptors impairs memory in the rat. Neuroreport 9:1407–1411. to synapses. Neuron 35:345–353. Jerusalinsky D, Kornisiuk E, Alfaro P, Quillfeldt J, Ferreira A, Rial VE, Bartus RT, Reginald L, Dean RL, Beer B, Lippa AS. 1982. The cholin- Durań R, Cervenãnsky C. 2000. Muscarinic toxins: novel pharmacologi- ergic hypothesis of geriatric memory dysfunction. Science 217:408–417. cal tools for the muscarinic cholinergic system. Toxicon 38:747–761. Bliss TV, Lomo T. 1973. Long-lasting potentiation of synaptic transmis- Kornisiuk E, Jerusalinsky D, Cervenãnsky C, Harvey AL. 1995. Binding sion in the dentate area of the anaesthetized rabbit following stimulation of muscarinic toxins MTx1 and MTx2 from the venom of the green of the perforant path. J Physiol 232:331–356. mamba Dendroaspis angusticeps to cloned human muscarinic cholinocep- Bonner TI, Buckley NJ, Young AC, Brann MR. 1987. Identification of tors. Toxicon 33:11–18. Corrigendum, Toxicon 33:1111. a family of muscarinic acetylcholine receptor genes. Science 237:527– Kornisiuk E, Sańchez G, Cervenãnsky C, Durań R, Jerusalinsky D. 532. 2001. Muscarinic toxins and their action on muscarinic inhibition of Bradford MM. 1976. A rapid and sensitive method for the quantitation adenylyl cyclase activity. J Neurochem 78(Suppl 1):BP18–BP37,183. of microgram quantities of protein utilizing the principle of protein-dye Leung LS, Shen B, Rajakumar N, Ma J. 2003. Cholinergic activity binding. Ann Biochem 72:248–254. enhances hippocampal long-term potentiation in CA1 during walking Bradley KN. 2000. Muscarinic toxins from the green mamba. Pharmacol in rats. J Neurosci 23:9297–9304. Ther 85:87–109. Levey AI, Edmunds SM, Koliatsos V, Wiley RG, Heilman CJ. 1995. Cobb S, Davies C. 2005. Cholinergic modulation of hippocampal cells Expression of m1–m4 muscarinic acetylcholine receptor proteins in rat and circuits [review]. J Physiol 562:81–88. hippocampus and regulation by cholinergic innervation. J Neurosci Descarries L, Gisiger V, Steriade M. 1997. Diffuse transmission by acetyl- 15:4077–4092. choline in the CNS. Prog Neurobiol 53:603–625. Felder CC, Porter AC, Skillman TL, Zhang L, Bymaster FP, Nathanson Li S, Cullen WK, Anwyl R, Rowan MJ. 2007. Muscarinic acetylcholine NM, Hamilton SE, Gomeza J, Wess J, McKinzie DL. 2001. Elucidating receptor-dependent induction of persistent synaptic enhancement in rat the role of muscarinic receptors in psychosis. Life Sci 68:2605–2613. hippocampus in vivo. Neuroscience 144:754–761. Fernańdez de Sevilla D, Nuń˜ez A, Birde M, Malinow R, BunõW. Lynch G, Rex CS, Gall CM. 2007. LTP consolidation: substrates, ex- 2008. Cholinergic-mediated IP3-receptor activation induces long-last- planatory power, and functional significance. Neuropharmacology ing synaptic enhancement in CA1 pyramidal neurons. J Neurosci 52:12–23. 28:1469–1478. Markram H, Segal M. 1990. Acetylcholine potentiates responses to N- Ferreira AR, Fu¨rstenau L, Blanco C, Kornisiuk E, Sańchez G, Daroit D, methyl-D-aspartate in the rat hippocampus. Neurosci Lett 113:62–65. Castro e Silva M, Cervenãnsky C, Jerusalinsky D, Quillfeldt JA. 2003. Max SI, Liang J-S, Purkenson SL, Potter LT. 1993. m4-Toxin, a selec- Role of hippocampal M1 and M4 muscarinic receptor subtypes in tive, reversible, allosteric antagonist of m4 muscarinic receptors. Soc memory consolidation in the rat. Pharmacol Biochem Behav 74:411– Neurosci Abstr 19:462. 415. Miyakawa T, Yamada M, Duttaroy A, Wess J. 2001. Hyperactivity and Fibiger HC. 1991. Cholinergic mechanisms in learning, memory and de- intact hippocampus-dependent learning in mice lacking the M1 musca- mentia: a review of recent evidence. Trends Neurosci 14:220–223. rinic acetylcholine receptor. J Neurosci 21:5239–5250.

Journal of Neuroscience Research 700 Sa´nchez et al.

Olianas M, Adem A, Karlsson E, Onali P. 1996. Rat striatal muscarinic Shirey JK, Xiang Z, Orton D, Brady AE, Johnson KA, Williams R, receptors coupled to the inhibition of adenylyl cyclase activity: potent Ayala JE, Rodriguez AL, Wess J, Weaver D, Niswender CM, Conn block by the selective m4 ligand muscarinic toxin 3 (MT3). Br J Phar- PJ. 2008. An allosteric potentiator of M4 mAChR modulates hippo- macol 118:283–288. campal synaptic transmission. Nat Chem Biol 4:42–50. Quirion R, Wilson A, Rowe W, Aubert I, Richard J, Doods H, Parent Sim JA, Griffith WH. 1996. Muscarinic inhibition of glutamatergic trans- A, White N, Meaney MJ. 1995. Facilitation of acetylcholine release missions onto rat magnocellular basal forebrain neurons in a thin-slice and cognitive performance by an M2-muscarinic receptor antagonist in preparation. Eur J Neurosci 8:880–891. aged memory-impaired rats. J Neurosci 15:1455–1462. Sokolov MV, Kleschevnikov AM. 1995. Atropine suppresses associative Rouse S, Marino M, Potter L, Conn P, Levey I. 1999. Muscarinic re- LTP in the CA1 region of rat hippocampal slices. Brain Res 672:281–284. ceptor subtypes involved in hippocampal circuits. Life Sci 64:501–509. Stoppini L, Buchs PA, Muller D. 1991. A simple method for organotypic Review. cultures of nervous tissue. J Neurosci Methods 37:173–182. Rowe WB, O’Donnell JP, Pearson D, Rose GM, Meaney MJ, Quirion Tai C, Kuzmiski JB, MacVicar BA. 2006. Muscarinic enhancement of R-type calcium currents in hippocampal CA1 pyramidal neurons. R. 2003. Long-term effects of BIBN-99, a selective muscarinic M2 receptor antagonist, on improving spatial memory performance in aged J Neurosci 26:6249–6258. cognitively impaired rats. Behav Brain Res 145:171–178. Vilaro´ MT, Mengod G, Palacios JM. 1993. Advances and limitations of the molecular neuroanatomy of cholinergic receptors: the example of Salamoni SD, Costa da Costa J, Palma MS, Konno K, Nihei K, Tavares multiple muscarinic receptors. Prog Brain Res 98:95–101. AA, de Abreu DS, Venturin GT, de Borba Cunha F, de Oliveira RM, Volpicelli LA, Levey AI. 2004. Muscarinic acetylcholine receptor sub- Breda RV. 2005. Antiepileptic effect of acylpolyaminetoxin JSTX-3 on types in cerebral cortex and hippocampus. Prog Brain Res 145:59–66. rat hippocampal CA1 neurons in vitro. Yajeya J, De La FA, Criado JM, Bajo V, Sanchez-Riolobos A, Heredia M. Santiago MP, Potter LT. 2001. Biotinylated m4-toxin demonstrates more 2000. Muscarinic agonist carbachol depresses excitatory synaptic transmis- M4 muscarinic receptor protein on direct than indirect striatal projec- sion in the rat basolateral amygdala in vitro. Synapse 38:151–160. tion neurons. Brain Res 894:12–20. Ye L, Qi J-S, Qiao J-T. 2001. Long-term potentiation in hippocampus Seeger T, Fedorova I, Zheng F, Miyakawa T, Koustova E, Gomeza J, of rats is enhanced by endogenous acetylcholine in a way that is inde- Basile AS, Alzheimer C, Wess J. 2004. M2 muscarinic acetylcholine re- pendent of N-methyl-D-aspartate receptors. Neurosci Lett 300:145– ceptor knock-out mice show deficits in behavioral flexibility, working 148. memory, and hippocampal plasticity. J Neurosci 24:10117–10127. Zhang W, Basile AS, Gomeza J, Volpicelli LA, Levey AI, Wess J. 2002. Shinoe T, Matsui M, Taketo MM, Manabe T. 2005. Modulation of syn- Characterization of central inhibitory muscarinic autoreceptors by the aptic plasticity by physiological activation of M1 muscarinic acetylcho- use of muscarinic acetylcholine receptor knock-out mice. J Neurosci line receptors in the mouse hippocampus. J Neurosci 25:11194–11200. 22:1709–1717.

Journal of Neuroscience Research